Stanford Advisors


All Publications


  • Regularization in multiplexed PET image reconstruction. Medical physics Foster, J. W., He, W., Chinn, G., Levin, C. S. 2026; 53 (11)

    Abstract

    Multiplexed PET (mPET) using one tracer labeled with a positron plus prompt gamma ray emitter is limited by low sensitivity to triple-photon coincidences (two annihilation photons and one prompt gamma photon), leading to high noise in the reconstructed images.The use of regularization has been previously shown to increase the image quality of single tracer PET images under low signal-to-noise ratio conditions. We will demonstrate the use of regularization in simultaneously acquired multi-tracer mPET data reconstruction.First, we present how a maximum a posteriori expectation maximization solution can be obtained for simultaneously acquired multi-isotope PET data. Next, we use a Monte Carlo simulation of an image quality phantom mPET acquisition with a pure positron (18F) and a positron plus prompt gamma (44Sc) emitter to evaluate the performance of different regularization methods.For a baseline regularization method to simultaneously reconstruct both activity distributions, we show that a relative difference penalty (RDP) provides the best tradeoff of the penalty functions tested in image quality and quantitative accuracy for a simulated hot sphere phantom compared to quadratic and log-cosh penalty functions, with a contrast-to-noise ratio (CNR) in 18F/44Sc of over 60% higher on average for the RDP compared to the quadratic penalty in the largest spheres. However, this regularization approach has up to 2x worse reconstructed inter-tracer crosstalk ratio compared to unregularized reconstruction. Finally, we present a further optimization by alternating reconstruction of the two individual tracer images with regularization to achieve a CNR in the largest 18F/44Sc spheres with an up to 35% improvement on average over the simultaneous RDP-based regularization reconstruction results described above, while maintaining a 10%-20% lower inter-tracer crosstalk ratio over the unregularized reconstruction.For the selected radionuclides, simulated system and phantom, we demonstrate proof-of-principle that using regularization applied to an alternating update form of EM achieves significantly higher CNR in both tracers compared to unregularized alternating update OSEM and regularized simultaneous OSEM, with significantly lower crosstalk in the prompt gamma emitter reconstruction than the regularized simultaneous OSEM.

    View details for DOI 10.1002/mp.70738

    View details for PubMedID 42949809

  • An optically coupled hybrid semi-monolithic PET detector for simplified DOI and position decoding. Physics in medicine and biology Zhao, Y., Yang, H., Huang, W., Zhang, X., Jiang, X., Niu, M., Zeng, X., He, W., Wei, Q., Gu, Z. 2026; 71 (16)

    Abstract

    Objective.Positron emission tomography (PET) detectors that encode depth-of-interaction (DOI) while achieving high timing and energy resolution are essential for long axial field-of-view (FOV) clinical PET systems. In this work, we propose an optically coupled hybrid semi-monolithic detector design, in which a pixelated scintillator array is coupled to a monolithic scintillator slab using optical adhesive. This design enables DOI capability while simplifying positioning calibration and preserving high timing and energy performance.Approach.Based on this concept, a prototype detector composed of a single-layer hybrid structure was fabricated and comprehensively evaluated. The prototype consists of aarray of pixelated lutetium-yttrium oxyorthosilicate (LYSO) crystals, each measuringwith a pitch of 3.2 mm, coupling via a high-refractive-index adhesive () to a monolithic LYSO slab of. The interface surfaces between the pixelated array and the monolithic slab were unpolished. The detector was read out using a TOFPET2 application-specific integrated circuit (ASIC).Main results.Experimental results show clearly separated pseudo-pixel event clusters in the flood image, enabling planar event positioning using a conventional energy-weighted average method. Using a 400-600 keV energy window, the detector achieved an average DOI resolution ofmm, an average energy resolution of, an intrinsic spatial resolution of 1.7 mm along the layer-stacking direction andmm along the pixelated array direction, and a coincidence timing resolution of 281 ps.Significance.These results demonstrate that the proposed optically coupled hybrid semi-monolithic detector balances detector performance (including DOI, timing and energy resolution) with calibration complexity. The design offers a promising detector solution for long axial FOV clinical PET systems.

    View details for DOI 10.1088/1361-6560/ae97d3

    View details for PubMedID 42575134

  • A system-level DOI discrimination method based on SSDA for a brain-dedicated DOI-PET scanner. Physics in medicine and biology Jiang, X., Zeng, X., Yang, H., Chen, Z., Huang, S., Huang, W., He, W., Niu, M., Gu, Z. 2026

    Abstract

    Depth-of-interaction (DOI) information is essential for improving the spatial resolution in positron emission tomography (PET) imaging. Existing deep learning-based DOI discrimination methods rely on side-irradiation measurements for each detector to obtain labeled data, making them impractical for system-level calibration.This study proposes a system-level DOI discrimination method based on semi-supervised domain adaptation (SSDA) for a brain-dedicated DOI-PET scanner. The method requires only side-irradiation data from a small number of detectors (even one) as labeled data, greatly reducing calibration time and labor. It was evaluated using a two-layer lutetium yttrium oxyorthosilicate (LYSO) detector array with different pixel sizes: the top layer is a 16 × 16 array of 1.53 × 1.53 × 5 mm³ crystals, and the bottom layer is an 8 × 8 array of 3 × 3 × 15 mm³ crystals. Performance was further evaluated on a PET prototype scanner comprising 72 detectors, through assessments of flood image quality and image spatial resolution of a 22Na point source positioned at different radial offsets.The method achieved an accuracy of 98.21% on a benchmark established using fully supervised learning, with flood image quality k of 4.80 ± 0.47 and 6.84 ± 0.53 for the top and bottom layers, and MAEk of 0.04 ± 0.03 and 0.07 ± 0.04, respectively. In the flood image quality evaluation of the 72-detector PET prototype scanner, it maintained stable and high performance comparable to fully supervised learning while using side-irradiation data from only one detector as labeled data, achieving k of 4.82 ± 0.55 and 6.65 ± 0.65 and identifying merely 0.80% misclassified crystal clusters. In the spatial resolution evaluation, the method achieved an average radial spatial resolution of 1.83 ± 0.45 mm for the two-level DOI configuration and 1.39 ± 0.20 mm for the four-level DOI configuration.With its low cost, high accuracy, and strong scalability, the proposed method provides an efficient and practical solution for system-level DOI calibration in next-generation high-performance brain-dedicated PET scanners.

    View details for DOI 10.1088/1361-6560/ae7eed

    View details for PubMedID 42309149

  • Influence of scintillation light confinement on depth-of-interaction measurement performance in a single-Ended readout PET detector. Physics in medicine and biology Shim, H. S., Cho, M. J., He, W., Lee, M. S., Levin, C. S., Lee, J. S. 2026

    Abstract

    Continuous depth-of-interaction (cDOI) detectors enable single-ended readout in positron emission tomography (PET) by encoding the interaction depth into the scintillation light distribution. This study presents a comprehensive performance optimization of a cDOI detector based on light distribution tailoring with crossed triangular-shaped reflectors by analyzing the effects of optical geometry based on various DOI decoding algorithms. Two parameters were systematically varied: the degree of light confinement through (1) optical segmentation and (2) crystal pitch adjustment. Five DOI decoding strategies-variance, max/sum ratio, Euclidean-distance classification, Gaussian and modified maximum-likelihood estimation, and artificial neural network (ANN) decoding-were adopted with 8×8 SiPM readout data. Results show that moderate segmentation (2×2 configuration) achieved the best DOI precision, yielding a 4.7 mm full width at half maximum (FWHM) and an ANN classification accuracy of 89 %. In the pitch study, 1.5 mm-pitch detector achieved better performance than 3.0 mm-pitch, indicating that increased optical interfaces allows more accurate encoding of the depth-dependent light distribution within the SiPM array. The ANN decoder consistently outperformed in DOI resolution compared to analytical and statistical methods by learning nonlinear spatial correlations among SiPM pixels. These findings highlight the coupled importance of optical geometry and data-driven decoding for achieving high DOI sensitivity in next-generation PET detector designs.

    View details for DOI 10.1088/1361-6560/ae6ead

    View details for PubMedID 42140287

  • Detector design for a high-resolution and high-sensitivity breast-dedicated multiplexed PET (mPET) scanner. Physics in medicine and biology He, W., Levin, C. S. 2026

    Abstract

    We are developing a high-resolution, high-sensitivity breast-dedicated PET scanner that can image the distribution of multiple tracers simultaneously, referred to as multiplexed PET (mPET). This requires the detector to have high intrinsic spatial resolution and detection efficiency, as well as resolve photon depth-of-interaction (DOI). Approach:The detector array design comprises a novel trapezoidal shape configuration with scintillation crystal rod elements of different lengths to enable adequate intrinsic detection efficiency for prompt gamma ray photons (>511 keV) in addition to 511 keV photons, which is also practical to fabricate and assemble. The LYSO element lengths range from 5 mm to 35 mm, with a 1.28 mm pitch, read out by a multi-pixel photon counter (MPPC) array with a 3 mm pixel size. A light-sharing approach, combined with unpolished crystal surfaces and top-side light guides, enables DOI encoding. A prototype detector module was experimentally evaluated for flood image/crystal element separation, energy, DOI, and detector timing resolution. Measurements were performed under both side-irradiation and top-irradiation setups to assess the detector's performance across different interaction positions. Main results:The prototype detector flood image successfully resolved crystal elements even for 35 mm length crystals. DOI resolutions in full width at half maximum (FWHM) measured 4-6 mm for crystal lengths ranging from 10 mm to 30 mm and 7-8 mm for the 35 mm crystal length. Energy resolutions varied from 13% to 28%, with degradation observed at positions farther from the MPPC array plane. Detector time resolution (DTR) values ranged from 330 ps to 660 ps.This work introduces a novel, single-ended readout, DOI-capable detector design that optimizes sensitivity while maintaining high spatial resolution for a breast-dedicated mPET system, utilizing a unique trapezoidal crystal array and light-sharing DOI encoding designs. Additionally, this design offers a scalable approach that can be adapted to other high-performance PET systems.

    View details for DOI 10.1088/1361-6560/ae50c9

    View details for PubMedID 41812326

  • Performance evaluation of S-PET, a compact, LYSO/BGO phoswich detector small animal PET prototype. Physics in medicine and biology Yang, H., He, W., Zhao, X., Li, G., Zhao, Y., Huang, W., Huang, Y., Niu, M., Liang, Y., Gu, Z. 2026; 71 (5)

    Abstract

    Objective.This work aims to evaluate the performance characteristics of S-positron emission tomography (PET), a compact, phoswich detector small animal PET prototype developed at Shenzhen Bay Laboratory (SZBL), in accordance with the National Electrical Manufacturers Association (NEMA) NU4-2008 protocol.Approach.The S-PET system comprises twenty flat-panel detectors arranged in two rings, with an inner diameter of 80 mm and an axial field of view (FOV) of 104.5 mm. Each detector adopts a phoswich depth of interaction design, consisting of two layers of scintillator crystal arrays, an oblique edge-cut glass lightguide, and two multi-pixel photon counter arrays. The front layer (annihilation photon entrance) is a 30 × 60 pixelated cerium-doped lutetium yttrium orthosilicate (LYSO) array (0.79 × 0.79 × 5 mm3per crystal), while the back layer (towards multi-pixel photon counters) is a 20 × 40 pixelated bismuth germanate (BGO) array (1.22 × 1.22 × 7.5 mm3per crystal). This phoswich design enables identification of most cross-layer crystal scatter (CLCS) events. Performance evaluation was conducted following NEMA NU4-2008 standards, including sensitivity, energy resolution, spatial resolution, noise equivalent count rate (NECR), scatter fraction (SF), image uniformity, spillover ratio, and recovery coefficient (RC).Main results.Within the 350-650 keV energy window, the peak absolute sensitivity at the FOV center was 14.9% (with CLCS events) and 11.9% (without CLCS events). The average system energy resolution was 10.7% ± 2.0% for LYSO and 28.1% ± 4.5% for BGO, derived from averaging individual crystal spectra. For three-dimensional ordered-subsets expectation maximization reconstructed images with spatially variant point spread function modeling, the spatial resolution of a point source in air ranged from 0.65 mm to 1.08 mm, with an average of 0.78 ± 0.12 mm across all measured locations. For the mouse-sized phantom, the peak NECR was 237.4 kcps at 11.9 MBq, with a SF of 13.0%; for the rat-sized phantom, these values were 110.3 kcps at 11.9 MBq and 23.3%, respectively. For the NEMA image quality phantom, the uniformity was 6.3%, spillover ratios in water-filled and air-filled cold region chambers were 11.1% and 7.9%, respectively, and RCs ranged from 0.36 to 0.89.Significance.The comprehensive performance results demonstrate that the S-PET prototype developed at SZBL possesses high spatial resolution and high sensitivity. These characteristics meet the requirements of preclinical research, particularly forin vivoimaging of small animals such as mice and rats.

    View details for DOI 10.1088/1361-6560/ae4b02

    View details for PubMedID 41747388

  • Optimization of TOF and DOI Performance in a Multi-Resolution Detector for Brain-Dedicated PET IEEE TRANSACTIONS ON RADIATION AND PLASMA MEDICAL SCIENCES Huang, S., Jiang, X., Chen, Z., Zeng, X., Huang, W., Yang, H., He, W., Niu, M., Wu, J., Wei, Q., Gu, Z. 2026; 10 (3): 380-390
  • Labeling of estradiol with a positron plus prompt gamma ray emitting radiometal for multiplexed positron emission tomography (mPET) imaging of breast cancer Cheung, P., Pandrala, M., He, W., Levin, C., Beinat, C. SOC NUCLEAR MEDICINE INC. 2025
  • A fast and accurate image reconstruction method for multiplexed PET (mPET) He, W., Foster, J., Levin, C. SOC NUCLEAR MEDICINE INC. 2025